LTP is triggered by Ca2+ entry through the NMDAR;subsequently Ca2+ activates calmodulin (CaM), which then activates CaMKII. Despite extensive studies demonstrating the pivotal role of CaMKII in LTP and memory, the mechanisms of its activation in living cells is not known. Intellectual Merit: The goal of the proposed work is to understand these mechanisms in quantitative detail. This requires methods to measure biochemical events in single spines near the limit of optical resolution and a sophisticated modeling framework for simulating these reactions. Because the experimental and computational methods were not previously available, this will be the first attempt to account for the measured activation of an enzyme in a living cell. Aim 1. Measurements will be made of critical quantitative properties of the system. The levels of free CaM will be measured using an optical reporter. This data, when analyzed using a computational modeling of the multiple equilibria involved, will yield the first estimate of free and bound CaM pools in neurons. The total concentration of Ng and CaMKII will also be measured. Aim 2. To model CaM activation requires information about the spatial/temporal gradients of Ca2+ in spines. 2-photon uncaging of glutamate will be used to activate NMDARs in a controlled way;the resulting Ca2+ elevation in the bulk spine cytoplasm will be measured under simplified conditions (see Aim 3). Ca2+ elevation in microdomains of the NMDAR is important for CaMKII activation (see below). To determine the elevation in microdomains of NR2A and NR2B, a stochastic model of NMDAR activation together with a modeling of Ca2+ diffusion and buffering will be developed to account for the measured bulk Ca2+ activation. This computational framework can then be used to estimate Ca2+ elevation in the microdomains. Aim 3. Using recently developed methods (fluorescence lifetime methodology (FLIM)), the time course of CaMKII activation in single spines will be measured. The role of microdomains in activation of the kinase will be examined. Our preliminary results suggest unexpected complexity: 1) both microdomain and bulk Ca2+ entry are required;2) phosphatases may influence kinase activation;3) phosphorylation of neurogranin, a protein that binds CaM, may vary over time (see Aim 4). To simplify the system, phosphatases will be inhibited and Ng modulation will be disabled. CaMKII activation (and Ca2+ elevation) will be measured under these simplified conditions. Computer simulations will then be used to predict how the elevation of Ca2+, as determined in Aim 2, leads to CaMKII activation. This prediction will be compared to the measured activation. This highly constrained framework can then be used to investigate different models of the processes involved (localization of reactions and the diffusional and unbinding processes that make CaM available). Once this simplified system is understood, additional experiments and computer simulations will be used to understand the more complex processes that normally regulate and modulate these steps. Aim 4. Neurogranin (Ng) is an abundant postsynaptic protein that binds CaM and may be important in controlling the CaM that is available to activate CaMKII. Moreover, there are modulatory processes that phosphorylate Ng and alter its ability to bind CaM. To determine whether such modulatory processes indeed affect plasticity, the effects of activating the metabotropic glutamate receptor will be studied. Previous work indicates that this leads to Ng phosphorylation and blocks LTP. To determine whether the Ng-mediated control of CaM mediates this effect on LTP, a Ng knockout mouse will be utilized. Cells will be transfected with a form of Ng that cannot be phosphorylated by PKC. If this blocks the effect of mGluR on LTP, it would demonstrate that the control of CaM by Ng phosphorylation can modulate plasticity. Broader Impact: To convey to the public progress in understanding memory, a large-scale mural illustrating the early steps in LTP will be mounted in a museum-like space at Brandeis University and on a website. As part of this grant, we will develop state-of-the-art software for Monte Carlo modeling of biochemical reactions. This software will be made available. The project will involve training, in part through the Posse Foundation minority program. There are also medical implications: understanding the early steps of synaptic plasticity may give insight into learning disorders. Ng, one of the proteins to be studied, is a risk gene for schizophrenia.